Endoplasmic reticulum-tethered transcription factor cAMP responsive element-binding protein, hepatocyte specific, regulates hepatic lipogenesis, fatty acid oxidation, and lipolysis upon metabolic stress in mice

Authors

  • Chunbin Zhang,

    1. Center for Molecular Medicine and Genetics,The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Guohui Wang,

    1. Center for Molecular Medicine and Genetics,The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Ze Zheng,

    1. Center for Molecular Medicine and Genetics,The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Krishna Rao Maddipati,

    1. Karmanos Cancer Institute, The Wayne State University School of Medicine, Detroit, MI
    2. Department of Pathology, The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Xuebao Zhang,

    1. Center for Molecular Medicine and Genetics,The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Gregory Dyson,

    1. Karmanos Cancer Institute, The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Paul Williams,

    1. Center for Molecular Medicine and Genetics,The Wayne State University School of Medicine, Detroit, MI
    Search for more papers by this author
  • Stephen A. Duncan,

    1. Department of Cell Biology, Neurobiology, and Anatomy, Medical College of Wisconsin, Milwaukee, WI
    Search for more papers by this author
  • Randal J. Kaufman,

    1. Neuroscience, Aging, and Stem Cell Research Center, Sanford-Burnham Medical Research Institute, La Jolla, CA
    Search for more papers by this author
  • Kezhong Zhang

    Corresponding author
    1. Center for Molecular Medicine and Genetics,The Wayne State University School of Medicine, Detroit, MI
    2. Department of Immunology and Microbiology,The Wayne State University School of Medicine, Detroit, MI
    3. Karmanos Cancer Institute, The Wayne State University School of Medicine, Detroit, MI
    • Center for Molecular Medicine and Genetics, Wayne State University School of Medicine, 540 East Canfield Avenue, Detroit, MI 48201===

    Search for more papers by this author
    • fax: 313-577-5218


  • Potential conflict of interest: Nothing to report.

  • Portions of this work were supported by American Heart Association grants 0635423Z and 09GRNT2280479 (to K.Z.), National Institutes of Health (NIH) grants DK090313 and ES017829 (to K.Z.), National Center for Research Resources grant S10RR027926 (to K.R.M.), and NIH grants HL057346, DK042394, and HL052173 (to R.J.K.).

Abstract

cAMP responsive element-binding protein, hepatocyte specific (CREBH), is a liver-specific transcription factor localized in the endoplasmic reticulum (ER) membrane. Our previous work demonstrated that CREBH is activated by ER stress or inflammatory stimuli to induce an acute-phase hepatic inflammation. Here, we demonstrate that CREBH is a key metabolic regulator of hepatic lipogenesis, fatty acid (FA) oxidation, and lipolysis under metabolic stress. Saturated FA, insulin signals, or an atherogenic high-fat diet can induce CREBH activation in the liver. Under the normal chow diet, CrebH knockout mice display a modest decrease in hepatic lipid contents, but an increase in plasma triglycerides (TGs). After having been fed an atherogenic high-fat (AHF) diet, massive accumulation of hepatic lipid metabolites and significant increase in plasma TG levels were observed in the CrebH knockout mice. Along with the hypertriglyceridemia phenotype, the CrebH null mice displayed significantly reduced body-weight gain, diminished abdominal fat, and increased nonalcoholic steatohepatitis activities under the AHF diet. Gene-expression analysis and chromatin-immunoprecipitation assay indicated that CREBH is required to activate the expression of the genes encoding functions involved in de novo lipogenesis, TG and cholesterol biosynthesis, FA elongation and oxidation, lipolysis, and lipid transport. Supporting the role of CREBH in lipogenesis and lipolysis, forced expression of an activated form of CREBH protein in the liver significantly increases accumulation of hepatic lipids, but reduces plasma TG levels in mice. Conclusion: All together, our study shows that CREBH plays a key role in maintaining lipid homeostasis by regulating the expression of the genes involved in hepatic lipogenesis, FA oxidation, and lipolysis under metabolic stress. The identification of CREBH as a stress-inducible metabolic regulator has important implications in the understanding and treatment of metabolic disease. (Hepatology 2012)

The liver plays a key role in controlling lipid homeostasis through regulation of lipid biosynthesis, oxidation, and transport. In response to different metabolic conditions, hepatic lipid metabolism is tightly and coordinately regulated by nuclear receptors, transcription factors, and cellular enzymes.1 Metabolic signals, such as fatty acids (FAs), glucose, insulin, or inflammatory cytokines, can regulate the activity and/or abundance of key transcription factors to modulate hepatic lipid metabolism.2, 3 Many hepatic transcription factors, such as sterol regulatory element-binding protein-1c (SREBP-1c), liver X receptor (LXRα), peroxisome proliferator-activated receptors (PPARs), and carbohydrate-responsive element-binding protein (ChREBP), have been identified as important regulators of gene expression involved in gluconeogenesis, lipogenesis, FA oxidation, and lipid transport.

The endoplasmic reticulum (ER) is an organelle that plays important roles in stress response and cell metabolism. ER-transmembrane–signaling molecules regulate lipid metabolism and/or glucose biosynthesis in response to a variety of stress stimuli.4-7 cAMP responsive element-binding protein, hepatocyte specific (CREBH), is an ER-transmembrane transcription factor of the CREB/ATF (activating transcription factor) family expressed in gastrointestinal tract tissues, including the liver, pyloric stomach, duodenum, and ileum.8-10 Previously, we demonstrated that CREBH is activated through regulated-intramembrane proteolysis (RIP) in response to ER stress, proinflammatory cytokines, or bacterial endotoxin lipopolysaccharide (LPS).11 Proinflammatory cytokines, such as tumor necrosis factor alpha (TNF-α), interleukin (IL)6, and IL1β, can induce both the expression and activation of CREBH in liver hepatocytes. Interestingly, CREBH activates the expression of genes encoding major liver acute-phase response proteins or phosphoenolpyruvate carboxykinase 1 (PCK1), but not those involved in the classical ER stress response.9, 11 More recent studies have shown that CREBH is inducible by free FA or fasting and is regulated by PPARα.7, 12, 13 Gain- and loss-of-function studies suggested that CREBH regulates hepatic gluconeogenesis by activating the expression of the Pck1 and glucose-6-phosphatase (G6Pase) genes in the liver.7

In this study, we show that CREBH is a key metabolic regulator of lipid homeostasis. Under conditions of metabolic stress, CREBH is activated to function as a transcription factor to activate the genetic programs required for lipogenesis, FA metabolism, and lipolysis. CREBH activity is required to maintain lipid equilibrium in the liver, plasma, and abdominal fat tissue. The identification of CREBH as a key hepatic lipid regulator has important implications in the understanding and treatment of metabolic diseases.

Abbreviations

ACC1, acetyl CoA carboxylase 1; AHF, atherogenic high fat; Apo, apolipoprotein; ATF6, activating transcription factor 6; BOH, 3-hydroxybutyric acid; ChIP, chromatin immunoprecipitation; ChREBP, carbohydrate-responsive element-binding protein; CoA, coenzyme A; CREBH, cAMP responsive element-binding protein, hepatocyte specific; CYP, cytochrome P450; DiHETEs, dihydroxy eicosatrienoic acids; Elovl, elongation of very-long-chain fatty acids protein; ER, endoplasmic reticulum; FAs, fatty acids; FADS, fatty acid desaturase; FASN, fatty acid synthase; G6P, glucose-6-phosphatase; GFP, green fluorescent protein; HDL, high-density lipoprotein; HETEs, hydroxyeicosatetraenoic acids; IL, interleukin; LCAT, lecithin-cholesterol acyltransferase; LDL, low-density lipoprotein; LXRα, liver X receptor; mRNA, messenger RNA; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; PA, palmitate; PBS, phosphate-buffered saline; PCK1, phosphoenolpyruvate carboxykinase 1; PGs, prostaglandins; PGC, PPARγ coactivator; PPARs, peroxisome proliferator-activated receptors; PUFA, polyunsaturated FA; RT-PCR, reverse-transcriptase polymerase chain reaction; SCD1, Stearoyl-CoA desaturase 1; SREBP-1c, sterol regulatory element-binding protein-1c; TGs, triglycerides; TM, tunicamycin; TNF-α, tumor necrosis factor alpha; TX, thromboxane; XBP1, X-box binding protein 1.

Materials and Methods

Mouse Experiments.

The CrebH knockout mice with exons 4-7 of the CrebH gene deleted were previously described.10 CrebH null and wild-type control mice on a C57Bl/6J background at 3 months of age were used for the metabolic diet study. The atherogenic high-fat (AHF) diet (Paigen diet; 16% fat, 41% carbohydrate, 1.25% cholesterol, and 0.5% sodium cholate by weight) was purchased from Harlan Laboratories, Inc. (TD.88051; Indianapolis, IN). All the animal experiments were approved by the Wayne State University Institutional Animal Care and Use Committee and were carried out under the institutional guidelines for ethical animal use.

Measurement of Mouse Lipid Contents.

Liver-tissue and blood-plasma samples were prepared from the mice under normal chow, after fasting, or after the AHF diet. To determine hepatic triglyceride (TG) levels, liver tissue was homogenized in phosphate-buffered saline (PBS), followed by centrifugation. The supernatant was mixed with 10% Triton-100 in PBS for TG measurement using a commercial kit (BioAssay Systems, Hayward, CA). Mouse blood-plasma samples were subjected to quantitative analyses of plasma TG, total cholesterol, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) as previously described.6 Levels of the ketone body metabolite, 3- hydroxybutyric acid, in the plasma were determined using commercial kits (BioAssay Systems).

Isolation of Primary Hepatocytes From the CrebH Null and Wild-Type Mice.

To isolate murine primary hepatocytes, livers from both CrebH null and wild-type mice were first perfused by Hank's balanced salt solution, supplemented with 8 mM of N-2-hydroxylethylpiperazine-N′-2-ethanesulfonic acid (pH 7.35) and 1 mM of sodium pyruvate, followed by collagenase digestion. Hepatocytes were pelleted by centrifugation at 50×g for 2 minutes, washed three times with Dulbecco's modified Eagle's media, and then seeded in the collagen-coated plates with Williams' E media.

For a full description of materials and methods used in this work, see Supporting Materials.

Results

Metabolic Stress Signals Induce CREBH Cleavage.

Previously, we demonstrated that activation of CREBH, indicated by its cleavage, can be induced by ER stress or inflammatory stimuli.11 To further explore pathophysiological challenges that trigger CREBH cleavage, we cultured a human hepatoma cell line, Huh-7, expressing full-length human CREBH protein fused with N-terminal three-tandem copies of FLAG epitope protein. The design of three copies of FLAG tag has been used to study stimuli-regulated mechanisms for cleavage of CREBH and another ER stress sensor, ATF6, by immunoblotting.11, 14 We exposed Huh-7 cells expressing flag-tagged full-length CREBH protein to a variety of biochemical or pathophysiological stimuli, including ER stress-eliciting reagent tunicamycin (TM), saturated FA palmitate (PA), and insulin. In the absence of challenges, hepatoma cells expressing full-length CREBH displayed basal levels of CREBH protein cleavage (Fig. 1A). This was likely the result of ER stress caused by overexpression of exogenous full-length CREBH. Consistent with our previous study, TM treatment increased CREBH cleavage (Fig. 1A). Among the pathophysiological stimuli tested, we found that insulin and PA could induce CREBH cleavage. To confirm this finding, we isolated murine primary hepatocytes and stimulated them with insulin and PA. Both insulin and PA could induce CREBH cleavage in a time-dependent manner (Fig. 1B,C). Additionally, we examined CREBH activation in the liver of wild-type mice fed an AHF diet, which is known to induce atherosclerosis and/or nonalcoholic fatty liver disease (NAFLD) in murine models.15, 16 Higher levels of cleaved CREBH protein were detected in the livers of the mice under the AHF diet, compared to that under the normal chow diet (Fig. 1D).

Figure 1.

Metabolic stress signals can activate CREBH. (A) HuH-7 cells were transfected with plasmid DNA vector expressing full-length human CREBH protein fused with N-terminal 3 copies of Flag tag. At 48 hours after transfection, cells were treated with PBS (control), TM (5 μg/mL), insulin (100 nM), or PA (10 μM) for 6 hours. The nontransfected cells were included as negative controls. Western blotting analysis was performed to detect CREBH cleavage by using an antiflag antibody. Ins, insulin; CREBH (f), full-length CREBH; CREBH (a), activated/cleaved CREBH. Levels of GAPDH were determined for protein-loading controls. (B and C) Primary hepatocytes isolated from wild-type mice were treated with insulin (100 nM) for 10, 30, 40, 60, and 360 minutes or PA (10 μM) for 0.5, 1, 2, 4, and 8 hours. Western blotting analysis was performed to detect endogenous CREBH cleavage using a polyclonal anti-CREBH antibody. (D) Western blotting analysis of CREBH cleavage in the liver tissue samples from wild-type mice under the normal chow (NC) or AHF diet for 6 months. For (A-D), the values below the gels represent the activated form of CREBH protein signal intensities after normalization to full-length CREBH signal intensities. (E) Quantitative real-time RT-PCR analysis of expression of the CrebH mRNA in mouse primary hepatocytes treated with TM (5 μg/mL), PA (10 μM), or insulin (100 nM) for 6 hours. Expression values were normalized to β-actin mRNA levels. Fold changes of mRNA are shown by comparing to that of nontreated control hepatocytes. Each bar denotes the mean ± SD (n = 3 samples per treatment). (F) Quantitative real-time RT-PCR analysis of expression of the CrebH mRNA in the liver of age-matched male mice under the normal chow diet, after 16 hours of fasting, or on the AHF diet for 6 months. Expression values were normalized to β-actin mRNA levels. Fold changes of mRNA levels are shown by comparing to 1 of the mice under the normal chow diet. Each bar denotes the mean ± SEM (n = 6 mice per group). GAPDH, glyceraldehyde 3-phosphate dehydrogenase; SD, standard deviation; SEM, standard error of the mean.

In addition to CREBH cleavage, we examined the expression of murine CrebH messenger RNA (mRNA) in response to ER stress or metabolic signals. Although TM, insulin, and PA could induce CREBH protein cleavage, they only marginally induced the expression of CrebH mRNA (Fig. 1E). Furthermore, we examined the expression of the CrebH gene in the liver of mice under different metabolic conditions. Expression of CrebH mRNA was significantly increased in response to 16-hour fasting, a condition that stimulates lipolysis, FA oxidation, and gluconeogenesis17 (Fig. 1F). Together, these data suggest that activation of CREBH is inducible by metabolic conditions, such as increased saturated FA, insulin signals, or the AHF diet.

Deletion of CrebH Decreases Hepatic Lipid Contents But Increases Plasma TG in Mice Fed Normal Chow.

To explore the pathophysiologic roles of CREBH, we bred CrebH knockout mice on C57Bl/6J strain background and studied them under different metabolic conditions. Although the CrebH null mice did not display any visible morphological or development defect,10, 11 they did exhibit a slight reduction in body weight, compared to the wild-type control mice (Fig. 2A). Further analyses revealed that body-fat mass in the CrebH null mice was reduced, compared to that in the control mice, under the normal chow diet (Fig. 2B). To explore the cause of the body-weight and fat-mass reductions, we first measured food consumption of the CrebH null mice when fed the normal chow diet. The CrebH null and wild-type control mice consumed similar amounts of food (Fig. 2C), thus excluding the possibility that food intake is a contributing factor to the body-weight or fat-mass reduction observed in the CrebH null mice. Based on these initial analyses, we suspected that the reduced fat mass and body weight reflected an imbalance in lipogenesis, FA oxidation, lipolysis, and/or lipid transport in the CrebH null mice.1, 3 Indeed, the CrebH null mice displayed reduced levels of hepatic TG, plasma cholesterol, HDLs, and LDLs, compared to the wild-type mice (Fig. 2D,E). Surprisingly, the levels of plasma TG in the CrebH null mice were higher than that in the control mice (Fig. 2D). To further evaluate the involvement of CREBH in maintaining lipid homeostasis, we fasted the CrebH null and wild-type control mice for 16 hours. Supporting the roles of CREBH in lipolysis, the CrebH null mice maintained much higher levels of plasma TG than the control mice under nutrient starvation, a condition that stimulates lipolysis (Fig. 2F). To determine whether the impaired lipid profile in the CrebH null mice was associated with FA oxidation, we measured levels of ketone body 3-hydroxybutyric acid (BOH), a product of FA oxidation in the plasma. After fasting, levels of BOH were slightly reduced in the CrebH null mice, compared to that of the control mice (Fig. 2F).

Figure 2.

CrebH null mice show a reduction in whole body fat mass, hepatic TG, and plasma cholesterol, but an increase in plasma TG. (A) Body weights of CrebH null and wild-type control mice (5 months old) on the normal chow diet. Each bar denotes the mean ± SEM (n = 6 mice per group). (B) Whole body fat mass of CrebH null and wild-type control mice (5 months old) on the normal chow diet. Each bar denotes the mean ± SEM (n = 6 mice per group). *P < 0.05. (C) Daily food intakes of CrebH null and wild-type mice (5 months old) under the normal chow diet. Daily food intake was determined based on the measurement of food consumption of CrebH null and control mice in 1 week. Each bar denotes the mean ± SEM (n = 8 wild-type mice or 7 knockout mice). (D) Levels of hepatic and plasma TG in CrebH null and wild-type mice (5 months old) under the normal chow diet. Each bar denotes the mean ± SEM (n = 6 mice per group). *P < 0.05. (E) Levels of plasma total cholesterol, HDL, and LDL in CrebH null and wild-type control mice (5 months old) under the normal chow diet. Each bar denotes the mean ± SEM (n = 6 mice per group). *P < 0.05; **P < 0.01. (F) Levels of plasma TG and BOH in CrebH null and wild-type mice after 16 hours of fasting. Each bar denotes the mean ± SEM (n = 6 mice per group). **P < 0.01.

CrebH Null Mice Develop Profound Hepatic Steatosis and Hypertriglyceridemia When Fed the AHF Diet.

To further delineate the role of CREBH in hepatic lipid metabolism, we fed the CrebH null and wild-type control mice with the AHF diet. After 6 months on the AHF diet, the body-weight gain in the CrebH null mice was much less than that in the control mice (Fig. 3A). Strikingly, the CrebH null mice developed enlarged fatty livers with pale color under the AHF diet (Fig. 3B). After the AHF diet feeding, the liver mass of the CrebH null mice was approximately 2 times of that of the wild-type control mice. Correlating with reduced body-weight gain and enlarged fatty livers, the sizes of abdominal fat tissues in the CrebH null mice were significantly smaller than those in the control mice fed the AHF diet (Fig. 3B). Additionally, we observed oily blood plasma with a milky color in the AHF diet-fed CrebH null mice (Fig. 3C).

Figure 3.

CrebH null mice display reduced body weight gain, massive steatosis, diminished abdominal fat, and hypertriglyceridemia under the AHF diet. (A) Photograph of representative CrebH null and wild-type control mice after the AHF diet for 6 months (left), and the body weights of CrebH null and wild-type control mice under the normal chow diet or after the AHF diet (right). Each bar denotes the mean ± SEM (n = 6 mice per group). **P < 0.01. (B) Liver and abdominal fat tissue of wild-type control and CrebH null mice were visualized in situ (left) and liver mass of wild-type and CrebH null mice after 6 months on the AHF diet (right). Each bar denotes the mean ± SEM (n = 6). **P < 0.01. (C) Photograph of blood-plasma samples from representative CrebH null and wild-type mice at 6 months after the AHF diet. (D) Oil Red O staining of cytosolic lipids in the liver tissue sections of representative CrebH null and wild-type control mice after the AHF diet for 6 months (magnification: 600×). (E and F) Levels of plasma TG, total cholesterol, HDL, and LDL, and levels of hepatic TG in CrebH null and wild-type control mice after the AHF diet for 6 months. Each bar denotes the mean ± SEM (n = 6). **P < 0.01. SEM, standard error of the mean.

To determine whether the phenotypes in the body weight, liver, and fat mass, as well as oily plasma of the CrebH null mice, are relevant to impaired lipid metabolism, we performed Oil Red O lipid staining to visualize the morphology and distribution of hepatic lipids in the liver tissue of the CrebH null mice after the AHF diet. The livers from the wild-type mice displayed macrovesicular steatosis characterized by engorgement of the hepatocytes with lipid droplets (Fig. 3D). In contrast, massive lipid contents, characterized by nonstructural lipid smears, were detected in the liver tissues of the AHF diet-fed CrebH null mice, indicating impaired hepatic lipid metabolism in the absence of CREBH. Furthermore, levels of plasma TG in the CrebH null mice were significantly higher than those in the control mice after the AHF diet (Fig. 3E), consistent with the oily plasma phenotype. However, levels of plasma lipoproteins, including total plasma cholesterol, HDL, and LDL, were slightly decreased in the CrebH null mice after the AHF diet (Fig. 3E). The plasma lipid profiles of the AHF diet-fed CrebH null mice were consistent with those observed with the CrebH null mice fed normal chow or after fasting (Fig. 2D-F), implicating a prominent hypertriglyceridemia phenotype caused by CrebH deletion. Additionally, levels of hepatic TGs in the CrebH null mice were much higher than those in the control mice after the AHF diet (Fig. 3F).

Deletion of CrebH Leads to Impaired Fatty Acid Metabolism and Severe Steatohepatitis in Mice After the AHF Diet.

To evaluate the abnormal lipid mass in the liver of the AHF-fed CrebH null mice, we performed lipidomic analysis of eicosanoids and docosanoids in the liver tissues of CrebH null and wild-type control mice fed the AHF diet. Eicosanoids and docosanoids are highly bioactive metabolites synthesized through the oxidation of 20- and 22-carbon polyunsaturated FA (PUFA), respectively.18 Supporting a role of CREBH in FA oxidation, levels of PUFA metabolites, including hydroxyeicosatetraenoic acids (HETEs), lipoxins, prostaglandins (PGs) (e.g., 15-keto PGE1 and 15-keto PGE2), hydroxy docosahexaenoic acids, and dihydroxy eicosatrienoic acids (DiHETEs), were significantly reduced in the liver tissues of the CrebH null mice fed the AHF diet, compared to those of wild-type mice (Fig. 4A; Supporting Fig. 1). In particular, the synthesis of epoxy FA catalyzed by cytochrome P450 (CYP) enzymes, including HETEs, epoxyeicosatrienoic acids, and DiHETEs, appeared to be defective in the CrebH null mice. The downstream metabolism of eicosanoids by the PG dehydrogenases inactivates the inflammatory lipid metabolites.18 Deficiency in this inactivation pathway was indicated by the significantly diminished levels of 15-keto PGE2, 15-keto PGE1, and 5-OxoETE in the CrebH null liver. Interestingly, metabolism of PUFA to the primary bioactive lipids, such as PGE2 and thromboxane (TX)A2 (measured as TXB2), was not significantly altered, suggesting a potential proinflammatory effect of CrebH deletion through a decrease in the inactivating metabolism, rather than increasing the inflammatory mediators.19 Together, these results implicate that CrebH deletion impairs FA metabolism in the liver, which may partially contribute to abnormal accumulation of hepatic lipid contents and liver inflammation in the AHF diet-fed CrebH null mice.

Figure 4.

Deletion of CrebH leads to impaired FA metabolism and profound NASH in mice under the AHF diet. (A) Lipidomic analysis of eicosanoids and docosanoids in the liver tissues of CrebH null and wild-type control mice after the AHF diet for 6 months. Levels of lipid metabolites that were decreased or increased in the CrebH null liver, compared to that in the wild-type control liver (P value cutoff was <0.05), are shown. Each bar denotes the mean ± SEM (n = 4 mice per group). (B) Histological examination (hematoxylin and eosion staining) of liver tissue sections of CrebH null and wild-type mice after the AHF diet for 6 months (magnification: 200×). (C) Sirius staining of collagen deposition in the liver tissue of wild-type and CrebH null mice after the AHF diet (magnification: 200×). (D) Histological scoring for NASH activities in the livers of the CrebH null and wild-type mice after the AHF diet for 6 months. Grade scores were calculated based on the scores of steatosis, hepatocyte ballooning, lobular and portal inflammation, and Mallory bodies. Stage scores were based on the liver fibrosis. Mean ± SEM (n = 4) values are shown. P values were calculated by the Mann-Whitney U test. SEM, standard error of the mean.

Next, we evaluated NAFLD-associated activities in the CrebH null mice. Significant liver inflammation and fibrosis in the AHF diet-fed CrebH null animals were evidenced by increased hepatocyte ballooning, lobular and portal inflammation, and Mallory bodies, as well as collagen deposition (Fig. 4B,C). Using a histological scoring system for NAFLD,20, 21 we revealed that the CrebH null mice, not the wild-type mice, developed profound nonalcoholic steatohepatitis (NASH) after the AHF diet (Fig. 4D). Supporting the histological analysis, levels of alanine aminotranferease and aspartate aminotransferase, the key indicators of hepatotoxicity, were increased in the CrebH null mice after the AHF diet (Supporting Fig. 2A,B). Moreover, the expression of two NASH-associated proinflammatory cytokines,22 TNF-α and IL6, was also significantly increased in the liver of the AHF diet-fed CrebH null mice (Supporting Fig. 2C,D). Additionally, we assessed the effects of CrebH deletion on glucose/insulin homeostasis. CrebH deletion led to decreased blood-glucose levels in the mice after overnight fasting, consistent with a previous study (Supporting Fig. 3A).7 We confirmed that expression of the gene encoding the gluconeogenesis regulator, PCK1, was defective in the CrebH null mice upon overnight fasting (Supporting Fig. 3B).7, 9 The glucose tolerance test showed that the CrebH null mice displayed similar patterns in decreasing exogenously injected glucose, compared to the control mice (Supporting Fig. 3C,D). However, the insulin tolerance test revealed that the AHF diet-fed CrebH null mice were not responsive to insulin in decreasing blood glucose (Supporting Fig. 3F), consistent with the increased NASH activities in the CrebH null mice after the AHF diet (Fig. 4D).

CREBH Is Required for Transcriptional Activation of Genes Involved in Lipogenesis, FA and Cholesterol Metabolism, and Lipolysis.

To understand the molecular basis underlying the lipid phenotypes observed with the CrebH null mice, we performed gene microarray analysis using total liver mRNA from the CrebH null and wild-type control mice fed either normal chow or the AHF diet. Through this analysis, we identified a subset of lipid metabolism-associated genes whose expression was down-regulated in the CrebH null mice fed normal chow or the AHF diet (Supporting Figs. 4 and 5). To verify the microarray analysis results, we performed quantitative real-time reverse-transcriptase polymerase chain reaction (RT-PCR) analysis and confirmed that deletion of CrebH in the liver decreased the expression of five groups of genes encoding functions critical for lipid metabolism (Fig. 5): (1) lipogenic regulators, including ChREBP, LXRα, PPARγ-coactivator (PGC)-1α, PGC-1β, and fat-specific protein 27 (Fig. 5A); (2) TG synthesis enzymes, including FA synthase (FASN), acetyl coenzyme A (CoA) carboxylase 1 (ACC1), ACC2, Stearoyl-CoA desaturase 1 (SCD1), and diacylglycerol acetyltransferase 2 (DGAT2) (Fig. 5B); (3) enzymes or regulators in lipolysis and lipid transport, including apolipoprotein (Apo) C2, Apo4, ApoA5, Apof, lipolysis-stimulated lipoprotein receptor, lecithin-cholesterol acyltransferase (LCAT), sterol carrier protein 2 (SCP2), and acyl-CoA thioesterase 4 (ACOT4) (Fig. 5C); (4) FA elongation enzymes, including elongation of very-long-chain FAs protein (Elovl) 2, Elovl5, Elovl6, and peroxisomal trans-2-enoyl-CoA reductase (PECR) (Fig. 5D); and (5) FA oxidation or cholesterol biosynthesis enzymes, including carnitine palmitoyltransferase 1A (CPT1a), CYP4a10, CYP4a14, CYP2b9, CYP2b13, FA desaturase 1 (FADS1), FADS2, palmitoyl-CoA oxidase 1 (ACOX1), PPARα, 24-dehydrocholesterol reductase (DHCR24), and long-chain-FA-CoA ligase 1 (ACS1) (Fig. 5E). The identified gene-expression profile is consistent with the phenotypes observed with the CrebH null mice. First, decreased expression of the lipogenic regulators and TG synthesis enzymes is consistent with reduced de novo lipogenesis observed in the CrebH null mice fed normal chow (Fig. 2). Second, decreased expression of enzymes or regulators in FA elongation and oxidation or cholesterol biosynthesis may be responsible, at least partially, for abnormal accumulation of hepatic lipid metabolites in the AHF-fed CrebH null mice (Figs. 3D and 4A). Finally, defective expression of enzymes required for lipolysis and lipid transport may account for hypertriglyceridemia, reduced fat mass and body-weight gain, and massive steatosis in the CrebH null mice fed the AHF diet or normal chow (Figs. 2 and 3).

Figure 5.

Deletion of CrebH leads to defective expression of genes involved in lipogenesis, FA and cholesterol metabolism, and lipolysis. Total RNAs were isolated from the liver tissues of CrebH null and wild-type control mice under the normal chow diet or after the AHF diet for 6 months and were subjected to quantitative real-time RT-PCR analysis of expression of the genes involved in hepatic lipid metabolism. Expression values were normalized to β-actin mRNA levels. Fold changes of mRNA levels are shown by comparing to 1 of the control mice under the normal chow diet or the AHF diet. Each bar denotes the mean ± SEM (n = 4). *P < 0.05; **P < 0.01. SEM, standard error of the mean.

To determine whether some or all of the lipid-associated genes that were down-regulated in the CrebH null mice are direct CREBH targets, we performed chromatin-immunoprecipitation (ChIP) experiments to test the potential of CREBH in binding to the promoter regions of the lipid-associated genes regulated by CREBH (Fig. 5). We expressed an activated form of CREBH with flag tag or a green fluorescent protein (GFP) control protein in primary hepatocytes isolated from CrebH null mice using an adenoviral-based expression system. The ChIP analysis indicated that CREBH could bind to the promoter regions of the genes encoding the following: TG synthesis enzymes, including ACC1, ACC2, and FASN; lipolysis coactivators, including ApoC2 and ApoA4; and enzymes involved in FA elongation, including Elovl2, Elov5, and Elovl6 (Fig. 6A,B). To assure CREBH-mediated transcriptional regulation of these genes, we performed gene-expression analysis with the animals overexpressing the activated CREBH in the liver. Quantitative real-time RT-PCR analysis indicated that the activated CREBH could dramatically up-regulate the expression of its target genes confirmed by ChIP analysis as well as other genes involved in lipogenesis, lipolysis, and FA metabolism (Fig. 6C-E). Additionally, we examined the expression of the proteins encoded by the CREBH-targeted genes in the animals overexpressing the activated CREBH or GFP. Western blotting analysis confirmed that levels of SCD1, PGC1, and ApoA4 proteins in the liver tissue or blood serum of the mice expressing the activated CREBH were significantly increased, compared to the mice expressing GFP (Fig. 6F). Together, these results suggested that CREBH could directly bind to, and activate the expression of, some of the key genes encoding functions involved in lipogenesis, FA metabolism, and lipolysis.

Figure 6.

CREBH binds to, and activates the expression of, the genes involved in hepatic lipid metabolism. (A and B) ChIP analysis of CREBH-binding activity to the promoter regions of the genes involved in hepatic lipid metabolism. Primary hepatocytes isolated from the CrebH null liver were infected with adenovirus expressing GFP or an activated CREBH protein with N-terminal fused flag tag. PCR was performed to identify potential CREBH-binding regions in the genes whose expression was defective in the CrebH null liver. Mock ChIP was included as a negative control (N-ctl). PCR reactions with the genomic DNA isolated from sonicated cell lysates were included as positive controls (P-ctl). (C-E) Gene-expression analysis in the livers of mice overexpressing the activated CREBH. Total RNAs from mice infected with adenovirus expressing GFP or activated CREBH for 72 hours were subjected to quantitative real-time RT-PCR analysis of gene expression. Expression values were normalized to β-actin mRNA levels. Fold changes of mRNA levels are shown by comparing to 1 of the mice overexpressing GFP. Each bar denotes the mean ± SEM (n = 3). **P < 0.01. (F) Western blotting analysis of protein levels for CREBH in the nuclear extract (NE), SCD1, and PGC1 in the liver tissue lysates, and ApoA4 in the blood plasma, of mice overexpressing GFP or activated CREBH. Levels of GAPDH and albumin were included as loading controls for liver tissue lysate and blood plasma samples, respectively. SEM, standard error of the mean; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

CREBH Activity Facilitates De Novo Lipogenesis and Lipolysis.

To verify the roles of CREBH in lipid metabolism, we performed functional analyses for CREBH using both in vitro and in vivo systems. First, we challenged primary hepatocytes from CrebH null and wild-type mice with ER stress-inducing reagent TM or the saturated fatty acid, PA. As a response to acute liver injury induced by TM or feeding with PA, lipogenesis and lipid deposition are increased in hepatocytes.6, 23 Without stress challenges, a modest reduction in cytosolic lipid contents, as characterized by Oil Red O staining, was detected in the CrebH null primary hepatocytes, compared to the wild-type hepatocytes (Fig. 7A). After 12 hours of incubation with TM or PA, wild-type hepatocytes developed severe steatosis. In contrast, only small amounts of cytosolic lipids were accumulated in the CrebH null hepatocytes after incubation with either TM or PA (Fig. 7A), thus confirming the involvement of CREBH in hepatic lipogenesis. Second, to verify the role of CREBH in lipolysis, we performed a fat tolerance test to track the ability of the CrebH null mice in clearance of exogenously infused TG. The CrebH null and control animals were fasted for 16 hours, followed by oral gavage of pure olive oil composed mainly of TGs. After dietary TG absorption, the CrebH null mice failed to decrease plasma TG levels (Fig. 7B). At 8 hours after olive-oil infusion, the wild-type mice decreased plasma TG levels to basal levels. In contrast, plasma TG levels in the knockout mice were kept as high as that after olive-oil absorption, suggesting the defect of plasma TG lipolysis in the absence of CREBH. Because the expression of ApoC2, a component of mature chylomicron, is decreased in the CrebH null mice, we wondered whether CrebH deletion could affect dietary lipid absorption in mice. Analyses of food uptake and dietary fat absorption suggested no significant differences in food ingestion, dietary lipid absorption, and excreted fecal lipids between the CrebH null and wild-type mice (Supporting Fig. 6A-C). This was further evidenced by the comparable levels of plasma ApoB48, the key indicator of chylomicron, detected in the CrebH null and wild-type mice (Supporting Fig. 6D).

Figure 7.

CREBH is required for lipogenesis and lipolysis. (A) Oil Red O staining of cytosolic lipids in CrebH null and wild-type control primary hepatocytes after the treatment of vehicle, TM (5 μg/mL), or PA (100 μM) for 12 hours (magnification: 400×). (B) Fat tolerance test for CrebH null and wild-type control mice. Mice were fasted for 16 hours, followed by oral gavage of pure olive oil at the dose of 12 μL/g body weight. Plasma TG levels were determined immediately before and at 2, 4, 6, and 8 hours after olive-oil loading. Mean ± SEM (n = 5) values were shown. SEM, standard error of the mean.

Having established the essential roles for CREBH in de novo lipogenesis and lipolysis, we next tested whether CREBH could regulate these lipid-associated processes by expressing an activated form of human CREBH in the mouse liver.11 The activated CREBH protein was overexpressed in mouse liver tissue through tail-vein infusion of adenovirus expressing the activated CREBH into wild-type mice. Oil Red O staining of liver tissue sections revealed a dramatic accumulation of lipid droplets in livers of mice overexpressing the activated CREBH protein (Fig. 8A). Consistently, levels of hepatic TGs were significantly increased in mice overexpressing the activated CREBH (Fig. 8B), supporting the role of CREBH in promoting de novo lipogenesis. In contrast, levels of plasma TGs were significantly decreased in mice expressing the activated form of CREBH, compared to that of control mice expressing GFP (Fig. 8C), confirming the role of CREBH in facilitating lipolysis.

Figure 8.

Forced expression of activated CREBH in the liver increases lipogenesis and lipolysis. (A) Oil Red O staining of cytosolic lipids in the liver tissue sections of mice overexpressing GFP or the activated form of CREBH (magnification: 400×). Frozen liver tissue sections were prepared from mice at 72 hours after infection with adenovirus expressing GFP or activated CREBH through tail-vein injection. (B and C) Levels of hepatic TG and plasma TG in the livers of mice overexpressing GFP or activated CREBH. Each bar denotes the mean ± SEM (n = 3). **P < 0.01. SEM, standard error of the mean. (D) Working model for the role of CREBH in regulating lipid homeostasis under metabolic stress.

Discussion

In this study, we demonstrated that the ER-resident transcription factor, CREBH, regulates multiple lipid-associated pathways, including lipogenesis, FA and cholesterol metabolism, and lipolysis, to maintain lipid homeostasis under metabolic stress conditions (Fig. 8D). With CrebH knockout mice, we identified complex lipid phenotypes caused by CrebH deletion. Under the normal chow, CrebH null mice displayed reduced lipogenesis, as evidenced by lower levels of hepatic TG and plasma cholesterols, reduced whole body fat mass, and decreased expression of regulators and enzymes required for TG and cholesterol biosynthesis (Figs. 2 and 5). Simultaneously, CrebH deletion caused a defect in TG lipolysis, because CrebH null mice displayed higher levels of plasma TG and reduced expression of key activators of lipolysis, compared to wild-type control mice (Figs. 2D and 5C). The involvement of CREBH in maintaining lipid homeostasis was further evidenced by the phenotypes and gene-expression profile in CrebH null mice fed the AHF diet. AHF-fed CrebH null mice displayed hypertriglyceridemia, severe NASH, reduced body-weight gain, diminished abdominal fat, as well as insulin resistance (Figs. 3 and 4; Supporting Fig. 3). Lipid staining and lipidomic analysis revealed that the enlarged livers of CrebH null mice accumulated nonstructural lipid mass and significantly altered the profile of lipid metabolites (Figs. 3D and 4A). This phenotype is likely caused by the defects in FA and cholesterol metabolism in the absence of CREBH. In support of the role of CREBH in promoting lipogenesis and lipolysis, CrebH null primary hepatocytes showed the defect in accumulating cytosolic lipid droplets in response to stress or injuries (Fig. 7A), and CrebH null mice failed to decrease exogenously infused TG in the bloodstream (Fig. 7B). Indeed, during the review process of the manuscript for this article, a report was published that confirmed the requirement of CREBH for lipolysis.24

Among the list of the CREBH target genes, ApoC2 is a key activator of lipolysis. In response to nutrient overload, ApoC2 is secreted into plasma, where it activates lipoprotein lipase to hydrolyze TG, thus providing a source of free FA for cells. Mutations in this gene cause hyperlipoproteinemia, characterized by hypertriglyceridemia and xanthomas.25 Defective expression of ApoC2, and two other lipolysis coactivators, LCAT and ApoA4, likely accounts for hypertriglyceridemia and diminished peripheral fat mass observed in CrebH null mice, especially after the AHF diet. Because ApoC2 is a component of mature chylomicron,26 the vehicle of dietary lipids in the bloodstream, we suspected the ApoC2 defect may affect dietary lipid absorption in CrebH null mice. However, our analyses suggested no significant differences in food consumption and dietary lipid absorption between CrebH null and wild-type mice (Fig. 2C; Supporting Fig. 6A-C). Apparently, decreased expression of ApoC2 does not affect nascent chylomicron delivery of dietary TG from the small intestine in CrebH null mice. Instead, the ApoC2 defect leads to a failure in triggering the activity of lipoprotein lipase to break down dietary TG in the bloodstream. As a consequence, chylomicron remnants, in which the dietary TG store is not completely distributed, will be retained in the liver. This may partially contribute to hepatic TG accumulation in the liver of AHF diet-fed CrebH null animals. It should be noted that CrebH null mice exhibited outwardly opposite phenotypes in hepatic lipid accumulation upon the normal chow or AHF diet (Figs. 2 and 3). Under the normal chow diet, both lipogenesis and lipolysis are crucial in determining hepatic and plasma TG levels, and therefore, the roles of CREBH in both lipogenesis and lipolysis are visible, because CrebH null mice displayed lower hepatic TG levels, but higher plasma TG levels, compared to control mice (Fig. 2). However, under the AHF diet, FA elongation and oxidation, lipolysis, and lipid transport, but not de novo lipogenesis, are prevalent programs that determine lipid levels in the liver and other periphery organs.2, 3 The defects of FA oxidation, cholesterol metabolism, and lipolysis in CrebH null mice likely overrode the defect of de novo lipogenesis, leading to profound hepatic steatosis and hypertriglyceridemia in CrebH null mice after the AHF diet (Fig. 3).

In summary, our work implicates that CREBH plays key roles in maintaining lipid homeostasis by regulating multiple key metabolic pathways in the liver (Fig. 8D). Disruption of CREBH activity leads to complex lipid-associated phenotypes, including hepatic steatosis, hypertriglyceridemia, and reduction in fat mass and body weight. In contrast, overexpression of the activated CREBH in the liver can significantly increase de novo hepatic lipogenesis, but reduce plasma TG levels simultaneously. These findings may have an important implication in pharmaceutical intervention toward controlling lipid homeostasis by modulating CREBH activity. Our work has also raised many immediate, important questions. For example, what is the precise mechanism by which CREBH controls the synthesis of lipogenic enzymes or regulators? Does CREBH act alone or in partnership with other known transcription factors, such as SREBP, ChREBP, and LXRα? Additionally, other ER stress-induced transcription factors, such as ATF6 and X-box binding protein 1 (XBP1), also regulate hepatic lipid homeostasis.5, 27 It is interesting to elucidate whether CREBH could function with ATF6 and/or XBP1 in regulating lipid homeostasis. All these questions deserve future research effort on CREBH.

Acknowledgements

The authors thank Drs. James Granneman, Todd Leff, and Emilio Mottillo for their critical comments and technical help.

Ancillary